#include "splitmatrix.h"
#include "phylotree.h"
+#include "distancecommand.h"
+#include "seqsummarycommand.h"
/***********************************************************************/
taxFile = tax;
large = l;
}
+/***********************************************************************/
+
+SplitMatrix::SplitMatrix(string ffile, string name, string tax, float c, float cu, string t, int p, string output){
+ m = MothurOut::getInstance();
+ fastafile = ffile;
+ namefile = name;
+ taxFile = tax;
+ cutoff = c; //tax level cutoff
+ distCutoff = cu; //for fasta method if you are creating distance matrix you need a cutoff for that
+ method = t;
+ processors = p;
+ outputDir = output;
+}
/***********************************************************************/
if (method == "distance") {
splitDistance();
- }else if (method == "classify") {
+ }else if ((method == "classify") || (method == "fasta")) {
splitClassify();
}else {
m->mothurOut("Unknown splitting method, aborting split."); m->mothurOutEndLine();
if (large) { splitDistanceLarge(); }
else { splitDistanceRAM(); }
+
+ return 0;
}
catch(exception& e) {
int SplitMatrix::splitClassify(){
try {
cutoff = int(cutoff);
-
+
map<string, int> seqGroup;
map<string, int>::iterator it;
map<string, int>::iterator it2;
PhyloTree* phylo = new PhyloTree();
ifstream in;
- openInputFile(taxFile, in);
+ m->openInputFile(taxFile, in);
//read in users taxonomy file and add sequences to tree
string seqname, tax;
while(!in.eof()){
- in >> seqname >> tax; gobble(in);
-
+ in >> seqname >> tax; m->gobble(in);
phylo->addSeqToTree(seqname, tax);
}
in.close();
//make sure the cutoff is not greater than maxlevel
if (cutoff > phylo->getMaxLevel()) { m->mothurOut("splitcutoff is greater than the longest taxonomy, using " + toString(phylo->getMaxLevel())); m->mothurOutEndLine(); cutoff = phylo->getMaxLevel(); }
-
+
//for each node in tree
for (int i = 0; i < phylo->getNumNodes(); i++) {
//is this node within the cutoff
TaxNode taxon = phylo->get(i);
-
+
if (taxon.level == cutoff) {//if yes, then create group containing this nodes sequences
if (taxon.accessions.size() > 1) { //if this taxon just has one seq its a singleton
for (int j = 0; j < taxon.accessions.size(); j++) {
}
}
}
+
+ delete phylo;
+
+ if (method == "classify") {
+ splitDistanceFileByTax(seqGroup, numGroups);
+ }else {
+ createDistanceFilesFromTax(seqGroup, numGroups);
+ }
+
+ return 0;
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "SplitMatrix", "splitClassify");
+ exit(1);
+ }
+}
+/***********************************************************************/
+int SplitMatrix::createDistanceFilesFromTax(map<string, int>& seqGroup, int numGroups){
+ try {
+ map<string, int> copyGroups = seqGroup;
+ map<string, int>::iterator it;
+ set<string> names;
+
+ for (int i = 0; i < numGroups; i++) { //remove old temp files, just in case
+ remove((fastafile + "." + toString(i) + ".temp").c_str());
+ }
+
+ ifstream in;
+ m->openInputFile(fastafile, in);
+
+ //parse fastafile
+ ofstream outFile;
+ while (!in.eof()) {
+ Sequence query(in); m->gobble(in);
+ if (query.getName() != "") {
+
+ it = seqGroup.find(query.getName());
+
+ //save names in case no namefile is given
+ if (namefile == "") { names.insert(query.getName()); }
+
+ if (it != seqGroup.end()) { //not singleton
+ m->openOutputFileAppend((fastafile + "." + toString(it->second) + ".temp"), outFile);
+ query.printSequence(outFile);
+ outFile.close();
+
+ copyGroups.erase(query.getName());
+ }
+ }
+ }
+ in.close();
+
+ //warn about sequence in groups that are not in fasta file
+ for(it = copyGroups.begin(); it != copyGroups.end(); it++) {
+ m->mothurOut("ERROR: " + it->first + " is missing from your fastafile. This could happen if your taxonomy file is not unique and your fastafile is, or it could indicate and error."); m->mothurOutEndLine();
+ exit(1);
+ }
+
+ copyGroups.clear();
+
+ //process each distance file
+ for (int i = 0; i < numGroups; i++) {
+
+ string options = "fasta=" + (fastafile + "." + toString(i) + ".temp") + ", processors=" + toString(processors) + ", cutoff=" + toString(distCutoff);
+
+ Command* command = new DistanceCommand(options);
+
+ command->execute();
+ delete command;
+
+ remove((fastafile + "." + toString(i) + ".temp").c_str());
+
+ //remove old names files just in case
+ remove((namefile + "." + toString(i) + ".temp").c_str());
+ }
+
+ singleton = namefile + ".extra.temp";
+ ofstream remainingNames;
+ m->openOutputFile(singleton, remainingNames);
+
+ bool wroteExtra = false;
+
+ ifstream bigNameFile;
+ m->openInputFile(namefile, bigNameFile);
+
+ string name, nameList;
+ while(!bigNameFile.eof()){
+ bigNameFile >> name >> nameList; m->gobble(bigNameFile);
+
+ //did this sequence get assigned a group
+ it = seqGroup.find(name);
+
+ if (it != seqGroup.end()) {
+ m->openOutputFileAppend((namefile + "." + toString(it->second) + ".temp"), outFile);
+ outFile << name << '\t' << nameList << endl;
+ outFile.close();
+ }else{
+ wroteExtra = true;
+ remainingNames << name << '\t' << nameList << endl;
+ }
+ }
+ bigNameFile.close();
+
+ for(int i=0;i<numGroups;i++){
+ string tempNameFile = namefile + "." + toString(i) + ".temp";
+ if (outputDir == "") { outputDir = m->hasPath(fastafile); }
+ string tempDistFile = outputDir + m->getRootName(m->getSimpleName((fastafile + "." + toString(i) + ".temp"))) + "dist";
+
+ //if there are valid distances
+ ifstream fileHandle;
+ fileHandle.open(tempDistFile.c_str());
+ if(fileHandle) {
+ m->gobble(fileHandle);
+ if (!fileHandle.eof()) { //check for blank file - this could occur if all dists in group are above cutoff
+ map<string, string> temp;
+ temp[tempDistFile] = tempNameFile;
+ dists.push_back(temp);
+ }else {
+ ifstream in;
+ m->openInputFile(tempNameFile, in);
+
+ while(!in.eof()) {
+ in >> name >> nameList; m->gobble(in);
+ wroteExtra = true;
+ remainingNames << name << '\t' << nameList << endl;
+ }
+ in.close();
+ remove(tempNameFile.c_str());
+ }
+ }
+ fileHandle.close();
+ }
+
+ remainingNames.close();
+ if (!wroteExtra) {
+ remove(singleton.c_str());
+ singleton = "none";
+ }
+ if (m->control_pressed) { for (int i = 0; i < dists.size(); i++) { remove((dists[i].begin()->first).c_str()); remove((dists[i].begin()->second).c_str()); } dists.clear(); }
+
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "SplitMatrix", "createDistanceFilesFromTax");
+ exit(1);
+ }
+}
+/***********************************************************************/
+int SplitMatrix::splitDistanceFileByTax(map<string, int>& seqGroup, int numGroups){
+ try {
+ map<string, int>::iterator it;
+ map<string, int>::iterator it2;
+
ifstream dFile;
- openInputFile(distFile, dFile);
+ m->openInputFile(distFile, dFile);
ofstream outFile;
for (int i = 0; i < numGroups; i++) { //remove old temp files, just in case
remove((distFile + "." + toString(i) + ".temp").c_str());
}
-
//for buffering the io to improve speed
//allow for 10 dists to be stored, then output.
vector<string> outputs; outputs.resize(numGroups, "");
vector<int> numOutputs; numOutputs.resize(numGroups, 0);
+ //you can have a group made, but their may be no distances in the file for this group if the taxonomy file and distance file don't match
+ //this can occur if we have converted the phylip to column, since we reduce the size at that step by using the cutoff value
+ vector<bool> validDistances; validDistances.resize(numGroups, false);
+
//for each distance
while(dFile){
string seqA, seqB;
if (m->control_pressed) { dFile.close(); for (int i = 0; i < numGroups; i++) { remove((distFile + "." + toString(i) + ".temp").c_str()); } }
- dFile >> seqA >> seqB >> dist; gobble(dFile);
+ dFile >> seqA >> seqB >> dist; m->gobble(dFile);
//if both sequences are in the same group then they are within the cutoff
it = seqGroup.find(seqA);
if ((it != seqGroup.end()) && (it2 != seqGroup.end())) { //they are both not singletons
if (it->second == it2->second) { //they are from the same group so add the distance
- if (numOutputs[it->second] > 10) {
- openOutputFileAppend((distFile + "." + toString(it->second) + ".temp"), outFile);
+ if (numOutputs[it->second] > 30) {
+ m->openOutputFileAppend((distFile + "." + toString(it->second) + ".temp"), outFile);
outFile << outputs[it->second] << seqA << '\t' << seqB << '\t' << dist << endl;
outFile.close();
outputs[it->second] = "";
numOutputs[it->second] = 0;
+ validDistances[it->second] = true;
}else{
outputs[it->second] += seqA + '\t' + seqB + '\t' + toString(dist) + '\n';
numOutputs[it->second]++;
remove((namefile + "." + toString(i) + ".temp").c_str());
//write out any remaining buffers
- if (numOutputs[it->second] > 0) {
- openOutputFileAppend((distFile + "." + toString(i) + ".temp"), outFile);
+ if (numOutputs[i] > 0) {
+ m->openOutputFileAppend((distFile + "." + toString(i) + ".temp"), outFile);
outFile << outputs[i];
outFile.close();
outputs[i] = "";
numOutputs[i] = 0;
+ validDistances[i] = true;
}
}
ifstream bigNameFile;
- openInputFile(namefile, bigNameFile);
+ m->openInputFile(namefile, bigNameFile);
singleton = namefile + ".extra.temp";
ofstream remainingNames;
- openOutputFile(singleton, remainingNames);
+ m->openOutputFile(singleton, remainingNames);
bool wroteExtra = false;
string name, nameList;
while(!bigNameFile.eof()){
- bigNameFile >> name >> nameList; gobble(bigNameFile);
+ bigNameFile >> name >> nameList; m->gobble(bigNameFile);
//did this sequence get assigned a group
it = seqGroup.find(name);
if (it != seqGroup.end()) {
- openOutputFileAppend((namefile + "." + toString(it->second) + ".temp"), outFile);
+ m->openOutputFileAppend((namefile + "." + toString(it->second) + ".temp"), outFile);
outFile << name << '\t' << nameList << endl;
outFile.close();
}else{
}
}
bigNameFile.close();
- remainingNames.close();
-
- if (!wroteExtra) {
- remove(singleton.c_str());
- singleton = "none";
- }
-
+
for(int i=0;i<numGroups;i++){
string tempNameFile = namefile + "." + toString(i) + ".temp";
string tempDistFile = distFile + "." + toString(i) + ".temp";
+
+ //if there are valid distances
+ if (validDistances[i]) {
+ map<string, string> temp;
+ temp[tempDistFile] = tempNameFile;
+ dists.push_back(temp);
+ }else{
+ ifstream in;
+ m->openInputFile(tempNameFile, in);
- map<string, string> temp;
- temp[tempDistFile] = tempNameFile;
- dists.push_back(temp);
+ while(!in.eof()) {
+ in >> name >> nameList; m->gobble(in);
+ wroteExtra = true;
+ remainingNames << name << '\t' << nameList << endl;
+ }
+ in.close();
+ remove(tempNameFile.c_str());
+ }
}
+ remainingNames.close();
+
+ if (!wroteExtra) {
+ remove(singleton.c_str());
+ singleton = "none";
+ }
+
if (m->control_pressed) {
for (int i = 0; i < dists.size(); i++) {
remove((dists[i].begin()->first).c_str());
}
return 0;
-
}
catch(exception& e) {
- m->errorOut(e, "SplitMatrix", "splitClassify");
+ m->errorOut(e, "SplitMatrix", "splitDistanceFileByTax");
exit(1);
}
}
ofstream outFile;
ifstream dFile;
- openInputFile(distFile, dFile);
+ m->openInputFile(distFile, dFile);
while(dFile){
string seqA, seqB;
}
}
}
- gobble(dFile);
+ m->gobble(dFile);
}
dFile.close();
while(bigNameFile){
bigNameFile >> name >> nameList;
nameMap[name] = nameList;
- gobble(bigNameFile);
+ m->gobble(bigNameFile);
}
bigNameFile.close();
int numGroups = 0;
ifstream dFile;
- openInputFile(distFile, dFile);
+ m->openInputFile(distFile, dFile);
while(dFile){
string seqA, seqB;
}
}
}
- gobble(dFile);
+ m->gobble(dFile);
}
dFile.close();